THE NEW ZEALANDERS A REPORT ON THEIR ASSETS AND DEBTS

Transcription

1 net THE worth OF NEW ZEALANDERS A REPORT ON THEIR ASSETS AND DEBTS

2 Liability statement Statistics New Zealand and the Retirement Commission give no warranty that the information or data supplied contains no errors. However, all care and diligence has been used in processing, analysing and extracting the information. Neither Statistics New Zealand nor the Retirement Commission shall be liable for any loss or damage suffered by the customer consequent upon the use directly, or indirectly, of the information supplied in this product. Catalogue number ISBN number Recommended retail price $35.00 (includes 12.5% GST and P&P)

3 page i Preface The need for quality data on the savings patterns of New Zealanders has long been acknowledged. Such information is vital in informing policy debate in areas as diverse as retirement provision, social equity analysis, consideration of microeconomic trends in consumption and asset prices, and broader macroeconomic implications for growth, balance of payments constraints and foreign investment patterns. The 2001 Household Savings Survey (HSS) conducted between August and November 2001 is the first major national survey of wealth to be conducted in New Zealand. This reflects in part the internationally recognised difficulties of collecting information in this area, particularly given the sensitive nature of the information sought, the wide range of asset and liability types involved and the inherent difficulties of assigning market values to these instruments. The Retirement Commission contracted Statistics New Zealand to undertake this national survey of net worth. The survey represents a major achievement, although all those involved acknowledge that it is only a starting point, capturing as it does a picture of the net worth of people living in New Zealand at a specific point in time. Many questions raised by the data provided will only be able to be answered if the survey is repeated, allowing the analysis of major trends over successive periods. Another milestone was the use of an electronic questionnaire to collect the survey data. This was the first Statistics New Zealand survey to equip interviewers with laptops and transfer the information collected back electronically via a secure internet link. This report presents a statistical overview of the net worth of the New Zealand population and looks at the distribution of net worth, both for couples and non-partnered individuals, and examines the impact of demographic factors such as age, ethnic group and family structure on savings patterns. Many people were involved in the HSS, from the development of the survey and introduction of the new interviewing technology to the fieldwork and analysis of the results. Foremost has been the co-operation and input from David Feslier and other staff of the Retirement Commission and of the Scoping Group. Within Statistics New Zealand, a number of former and current staff have contributed greatly to the undertaking of the Household Savings Survey including Tanya Randall, Sophia A Court, Terri Hendry, Diane Ramsay and Tas Papadopoulos. The contribution of all these people to the success of the survey is greatly appreciated. We would also like to acknowledge the authors of this report, Tanya Randall, Sophia A Court, and Karin Henshaw of Statistics New Zealand and independent contractor David Preston. We look forward now to interested parties making use of this rich new source of information on the net worth of people living in New Zealand. Brian Pink Government Statistician Colin Blair Retirement Commissioner

4 page ii

5 page 1 Contents Chapter one: Highlights 2 Chapter two: Introduction 4 Chapter three: Concepts and survey population statistics 8 Chapter four: The distribution of net worth 18 Chapter five: Ethnic groups 30 Chapter six: Education and employment 34 Chapter seven: Family structure 42 Chapter eight: Income 52 Chapter nine: Assets 58 Chapter ten: Māori assets 78 Chapter eleven: Trusts 84 Chapter twelve: Debt 90 Chapter thirteen: Student loans 102 Chapter fourteen: International comparison 112 Appendix one: Technical notes 124 Glossary 132 Modules and information collected 139 Table listing 146 Appendix two: Data quality: comparasion and imputation 148 Appendix three: Other sources of information on New Zealander s 152 net worth or saving This is an interactive PDF. If you click the mouse on the chapter titles on the contents page it will take you to the beginning of the selected chapter. On the last page of each chapter selecting the back arrow will take you back to the contents page and selecting the forwards arrow will take you to the first page of the following chapter.

6 page 2 chapter one Highlights The 2001 Household Savings Survey (HSS) estimated the total net worth of New Zealanders aged 18 years and over at $ billion. This was a total asset value of $ billion and a total debt value of $ billion. The total median net worth was $60,000 for total adults 1 and $68,600 for total economic units. 2 As net worth is accumulated over a lifetime, the distribution of net worth in New Zealand is closely related to age. In general, young adults had the lowest median net worth, while the middle age groups (50-64 year olds) had the highest median net worth. Those who were 65 or older had a substantially higher net worth than young adults, yet a lower net worth than the middle age groups. Median net worth was higher for the employed and self-employed than it was for the unemployed. By ethnic group, median net worth was highest for those who identified with the European/ Pākehā ethnic group. This was followed by those from the Asian ethnic group. The Māori and Pacific peoples ethnic groups had similar lower median net worth. The HSS showed that net worth was unevenly distributed across the population. When using a population of total economic units, 16 percent had negative net worth while 30 percent had net worth over $200,000. This also reflects the age of the population, with approximately 43 percent of economic units aged having negative net worth compared with only two percent of those over 65 years old. 1 The HSS interviewed non-partnered individuals and couples. To derive a total population of all adults the couples net worth is halved and apportioned to each partner. 2 Economic units is the straight summation of the population of non-partnered individuals and the population of couples much like one parent families and two parent families can be summed to a total population of families.

7 page 3 The uneven distribution of net worth across the population was more evident when dividing the population into net worth deciles. Of the total population of economic units, the bottom decile had a total net worth of -$3.303 billion while the top decile had a total net worth of $ billion. In the age groups over 35, the top 20 percent of the population held around 60 percent of the total net worth for these groups. The HSS found that overall non-partnered individuals and couples had an estimated $16 in debts for every $100 in assets. This estimate varied more widely between age groups. For example, student loans had a significant impact in the younger age groups. So while 37 percent of non-partnered individuals aged had student loans they made up 76 percent of non-partnered individuals with negative net worth. The most valuable asset people had was residential property. This accounted for approximately 43 percent of the total value of assets and included homes, rental property, holiday homes and other property. The largest type of debt was mortgage debt, which made up approximately 80 percent of the total value of debts. The most commonly held assets were bank deposits, motor vehicles and property. The most common debt was credit card debt.

8 page 4 chapter two Introduction The 2001 Household Savings Survey (HSS) was a cross-sectional nationwide survey that collected information on the current assets and debts of New Zealanders. This survey is the first of its kind in New Zealand and is expected to act as a benchmark against which the results of future surveys can be measured. It collected information on the level, composition, distribution and accumulation of net worth in the population. Background In 1992, the Taskforce on Private Provision for Retirement identified the need for regular information covering patterns of voluntary private savings and life cycle statistics on income, expenditure, and wealth. In its interim report in July 1997 the Periodic Report Group, charged with reviewing retirement income policies, expressed disappointment that with one exception there had been little change since 1992 in the quality of statistics available to inform the debate on individual and national savings levels and patterns. In its 1997 final report, The Periodic Report Group was pleased to note that since the interim report, the Retirement Commissioner and the Government Statistician had established a working group to assess the statistical requirements for future periodic reviews. They stated that good statistics on saving and net worth are essential if the saving issue is to be properly analysed and debated. The output of the working group was instrumental in the Retirement Commission gaining funding in the 1999 budget review, to improve statistics on the level and distribution of net worth in New Zealand households. As a result Statistics New Zealand was contracted to carry out the Household Savings Survey (HSS). Purpose of this report The purpose of this report is to present some key results from the HSS, along with a brief analysis of major trends and findings. This analysis is by no means exhaustive, but provides users with a stepping stone for their own enquiries. As the HSS is New Zealand s first net worth survey there is no basis for comparison over time. Further surveys will enable this. The appendix to this report includes technical notes and explanations of the key measurement concepts involved.

9 page 5 Net worth Other sources of data (such as those from financial institutions) relate to aggregate data and do not provide an insight into the distribution of assets and debts across the population. Analysis to date, including Statistics New Zealand s Household Economic Survey, has looked at the flow of savings and the difference between income and expenditure. In contrast, the HSS looked at the stock of savings the total value of assets less the total value of debts. This is net worth. At the micro level it is an estimate of the resources a non-partnered individual or couple would have if they cashed up their total assets and settled all their debts on the day they were interviewed. Total assets total debts = net worth Inherent to a survey of net worth is the concept of measuring the current market value of an asset. This measure relies on the respondent giving their best estimate of the value of their assets by taking into account factors such as valuation documents, purchase price, condition of the asset and the current market. Caveats The level of net worth held by an individual is strongly related to that individual s age. It is important to remember this when considering the distribution of net worth across an adult population. For example, it is quite possible for a society with total income and net worth equality over the course of a life cycle to still have the majority of net worth held by a small portion of that society, depending upon the age distribution of the population. For example, if 20 percent of the population is aged over 65 years, then it is likely that these 20 percent will hold the majority of total net worth as older people have higher levels of net worth. This peculiarity needs to be allowed for when drawing conclusions about the equality of net worth distribution. International comparisons, or comparisons across ethnic groups, will need to take into account the age distributions of the relevant populations. Without standardising by age, it is useful to compare groups within defined age bands. In this report 10-year age bands have been used. This is not ideal as, for instance, 25 and 34 year olds are likely to have quite different net worth characteristics. However, the survey design does not support a higher level of detail. This survey only provides a snapshot of net worth characteristics. It is important that the data is not read as a commentary on net worth accumulation over time. For example, while the current net worth status of today s year olds reflects the economic and cultural circumstances of that particular group, it cannot be assumed that future groups will replicate these characteristics when they reach this age. For this reason a longitudinal survey 1 or repeated cross-sectional surveys is needed so that aspects of net worth can be properly commented upon. Users Key users of the HSS data are expected to be those assessing and advising on people s preparedness for retirement. In addition, it is hoped that this information will feed into other areas of policy debate such as economic management and the impact of a nation s savings habits on investment patterns, asset prices and balance of payments trends. The impact of student loans is likely to be of great interest to policymakers, and there are many private sector groups that would like a better understanding of the nation s savings behaviour. 1 A longitudinal survey involves interviewing the same respondents selected in the first year, over a period of years. Repeated cross-sectional surveys on the other hand select a new sample to interview each time the survey is conducted.

10 page 6 The Household Savings Survey in brief The sample sample size: 5,374 interviews response rate: 74 percent The sample included a Māori booster sample and was made up of 2,392 non-partnered individuals and 2,982 couples. This was rated up to a total population of 930,900 nonpartnered individuals and 855,900 couples. Collection method Interviews were conducted in person using an electronic questionnaire. 2 If a respondent was part of a couple, the couple was interviewed as one unit. Information on all assets and debts, as well as demographic information, was collected. Information on assets and debts was only collected for the selected non-partnered individual and the couple, not for other family or household members. Key concepts Non-partnered individual a respondent who did not live with a partner, but may live with family (such as children or parents) or non-family members. Couples where the respondent who was selected to participate in the survey lived with their partner they were interviewed as a couple. Definition of a partner living with them was self-defined. Economic units the two populations (non-partnered individuals and couples) combined to form one population. Individual characteristics for couples the individual characteristics (such as age and ethnicity) given to the couple were the characteristics of the partner randomly selected to take part in the survey. Additional information The Net Worth of New Zealanders standard tables and technical notes contains over 50 standard tables with supporting definitions and technical notes on the survey methodology, collection and questionnaire. This report is available on the Statistics New Zealand website free of charge or by contacting Statistics New Zealand (see Contacts back cover). Tables of the relative sampling errors for the standard tables are also available by contacting Statistics New Zealand. 2 See Appendix 1 for definition.

11 page 7

12 page 8 chapter three Concepts and survey population statistics This chapter introduces some of the key concepts used in this report and provides a brief overview of the characteristics and structure of the population covered in the 2001 Household Savings Survey (HSS). These concepts are repeated throughout this report enabling chapters to be read as standalone documents. Concepts Definitions A full list of definitions can be found in appendix one. Respondent Is the one person aged 18 or older per household who was randomly selected to participate in the survey. If the respondent had a partner living with them and the couple was interviewed as one unit, the person selected is defined as the respondent in the couple. Non-partnered individuals Are respondents who were not living with a partner, but may have been living with family (such as children or parents) or non-family members. Couples Are where respondents said their partner lived in the same household. No attempt was made to define couples based on the time spent together. They included legally married, de-facto and same-sex relationships. Couples were interviewed jointly as one economic unit. Economic units For the purpose of analysis the two populations, non-partnered individuals and couples, were combined to form one total population of economic units (see analysis of economic units below for more detail). Concepts relating to analysis of couples Net worth of a couple Where the selected individual was part of a couple, information was collected about the couple as a whole. The value of net worth discussed throughout this report is the total of both partners in the couple.

13 page 9 Assigning individual characteristics to a couple In contrast, information collected about demographic characteristics (such as age and ethnicity) referred to the partner who had been randomly selected to take part in the survey (unless otherwise stated). This was considered the most practical way of analysing this type of data. For example, couples aged are those where the selected respondent was in this age group. The random selection meant there should be little bias. Other ways to analyse couples include assigning the characteristic of the highest income earner to the couple (this is used in the Canadian Survey of Financial Security 1 ) or using the head of the household (a technique adopted by the US Survey of Consumer Finances 2 ). However it is thought that these techniques may lead to a predominance of males. Further manipulation of the HSS data beyond the scope of this report could look at using different methods. Regardless of the method used, factors such as the partner s age and labour-force status will have some impact on the net worth situation. Where possible this report s analysis looks at the characteristics of the respondent and their partner. Analysis of economic units The couple s net worth represents the situation of two people whose net worth is interlinked. One would therefore expect the value of net worth to be at least double that of a nonpartnered individual. However, the unique differences between how couples and non-partnered individuals operate mean other factors need to be considered. It is more useful to look at the two groups as two separate populations. These can be combined to form a total population of economic units much like sole parents and two-parent families may be counted as total families in other surveys. Characteristics of the respondent and their partner The following characteristics were collected separately from each member of a couple: age, ethnic group, labour-force status, occupation and highest qualification. Despite the practical necessity of assigning a respondent s characteristic to a couple it is useful to look at the respondent and their partners characteristics to see how the two correlate. Age There was a tendency for both members of a couple to be within the same age group. For example, 71 percent of respondents aged had partners in the same age bracket. Where the respondent s partner was not in the same age group they were usually very close. Ethnic group Generally, there was a strong correlation between the respondent s ethnicity and their partner s. For example, of the respondents who identified with the European/Pākehā ethnic group, 94 percent had partners who identified with the same ethnic group. This was less evident with the Māori ethnic group, where only 57 percent of respondents of Māori ethnicity lived with partners of Māori ethnicity. Labour force status There was a relatively strong correlation for labour-force status. Where the respondent was employed, 83 percent of these respondents had partners who were also in employment. Where the respondent was not in the labour force, 61 percent of these respondents had partners who were also not in the labour force. 1, 2 See International comparison chapter

14 page 10 Highest qualification and occupation When looking at how people were partnered by highest qualification and occupation the correlation is not as strong as for the previous characteristics. For this reason these characteristics are only used on the population of non-partnered individuals, unless otherwise stated. Survey population statistics Socio-demographic characteristics of non-partnered individuals and couples (by characteristic of the respondent) Age Figure 3.1: Distribution of population by age non-partnered individuals couples % age group Figure 3.1 illustrates the different age distributions between the two populations. This is a key factor when looking at asset and debt ownership and the resulting net worth. As expected, non-partnered individuals were noticeably younger with nearly 50 percent aged between 18 and 34 (29 percent of whom were years old). In contrast, only 23 percent of couples were in the age range There was also a larger proportion of nonpartnered individuals over 75 than couples (10 percent compared with 5 percent). Meaningful analysis therefore needs to consider the impact of age. For reference throughout the report, figures 3.2 and 3.3 show the actual proportions of non-partnered individuals and couples within each age group. The age groups in figure 3.3 are often used when the sample size is too small to support a 10-year breakdown. Figure 3.2: Distribution of population by age Age group Non-partnered individuals Couples % % and over 2 0

15 page 11 Figure 3.3: Distribution of population by age Age group Non-partnered individuals Couples % % and over Note for all graphs and tables in the report, age groups are not shown where the figures were too small to be statistically significant. For example, a graph may not show a category for the age group. Ethnic group In the HSS questionnaire, people were asked to list all the ethnic groups they identified with. During the data processing stage, each respondent was allocated to a single ethnic group. This priority recording system, which was used in the 1996 Census, provides practical ease of analysis as each respondent appears only once in the data to sum to the total population. This data does not necessarily give a total count of people of a particular ethnic group (see Glossary, Appendix 1 for further information). Figure 3.4: Distribution of population by ethnic group non-partnered individuals couples % European/Pākehā Māori Pacific peoples Asian Other ethnic group The distribution of the population by ethnic group, as shown in figure 3.4, matches the proportions when comparing with 2001 Census data. It is useful to look at the different age structures within the ethnic groups as these can impact on other measures such as the distribution of net worth. The proportion of non-partnered individuals and couples over 65 is notably higher for those of European/Pākehā ethnicity. For example, 22 percent of non-partnered individuals of European/Pākehā ethnicity were 65 years of age or older while the corresponding figure for Māori was 6 percent. This can be seen in figure 3.5. Figure 3.6 demonstrates this for couples by adopting the age and ethnic group of the partner who had been selected to take part in the survey.

16 page 12 Figure 3.5: Age distribution of non-partnered individual s ethnic groups Age group European/Pākehā Māori Pacific peoples Asian % % % % and over Total Figure 3.6: Age distribution of couple s ethnic groups Age group European/Pākehā Māori Pacific peoples Asian % % % % and over Total Marital status Figure 3.7 shows the proportion of non-partnered individuals and individuals in couples in each legal marital status group. The concept of individuals in couples is different from the convention used in the rest of the report for couples because, in this case, each partner is counted separately. For example, one partner may have never married while the other may be separated. The majority of non-partnered individuals (60 percent) identified as having never married. A substantial proportion of non-partnered individuals (38 percent) were either separated, divorced or widowed.

17 page 13 Figure 3.7: Legal marital status by age Never Separated Age group Married married or divorced Widowed % % % % Non-partnered individuals and over Total Individuals in couples and over Total Note: indicates the count was negligible Of all couples, (not individuals in couples) 81 percent were legally married. The proportion of couples who were legally married increased with age. Twenty-seven percent of couples aged were legally married, this increased to 83 percent when the respondent was aged and reached 100 percent for couples aged 85 and over. Employment and education Labour force status Figure 3.8: Distribution by labour force status non-partnered individuals couples % employed unemployed not in the labour force labour force status The distribution of the population by labour-force status in figure 3.8, reflects labour-force figures found in other Statistics New Zealand surveys. The majority of the population (57 percent of non-partnered individuals and 71 percent of couples) was employed.

18 page 14 The next largest proportion of the population was categorised as not in the labour-force. The proportion of this group was higher for non-partnered individuals (39 percent) than for couples (28 percent). This reflects the higher proportions of older, retired people in the population of non-partnered individuals. Only a small percentage of the population (4 percent of non-partnered individuals and 1 percent of couples) was unemployed. Once again, distribution is affected by age with higher unemployment rates for youths whether nonpartnered or in a couple. Occupation Figure 3.9: Proportion of non-partnered individuals in each occupation type occupation legislators, administrators, managers professionals technicians & associate professionals clerks service & sales workers agriculture & fisheries workers trades workers plant & machinery operators & assemblers elementary occupations % Figure 3.9 shows that the largest proportion of employed non-partnered individuals were service and sales workers (19 percent). When grouping the occupation types in to the two more generalised groups of blue collar occupations (the bottom four occupations in figure 3.9) and white collar occupations (the top five occupations in figure 3.9), more nonpartnered individuals were employed in white collar occupations (71 percent) than blue collar occupations. Fifty-four percent of service and sales workers were aged These jobs are generally those with part-time and casual hours. Highest qualification Figure 3.10: Highest qualification of respondents non-partnered individuals couples % none school post-school degree other vocational highest qualification

19 page 15 Figure 3.10 shows the distribution of the populations by highest qualification. Over half of non-partnered individuals and couples gave either no qualification or school qualification as their highest completed qualification. Twelve percent of non-partnered individuals and 15 percent of couples said a degree was their highest completed qualification. This included both under-graduate and post-graduate degrees. Children and income Children Figure 3.11: Number of children and proportion of population with dependent children and children ever had Dependent children Total children ever had Number Non-partnered Non-partnered of children individuals Couples individuals Couples % % % % or or more The HSS used the standard definition of a dependent child. A dependent child was one that was living in the same household as the respondent, was aged under 18 years of age and who was not employed full time. Figure 3.11 shows that the majority of non-partnered individuals and couples had no dependent children (83 percent and 57 percent respectively). There were more couples with dependent children than non-partnered individuals. The proportion of non-partnered individuals who have never had any children was much higher than the proportion of couples who have never had any children. There was also a higher proportion of couples who had had three or more children. Main source of income Figure 3.12 shows the distribution of the population by main source of income. As expected, the majority of the non-partnered individual and couple populations gave wages and salaries as their main source of income (51 percent and 61 percent respectively). The second largest group for non-partnered individuals was those who had other income support as their main source of income (22 percent) 3.The second largest group for couples was those who had self-employment as their main source of income. 3 This covers all government benefits and allowances such as student allowances, family support and ACC earnings-related payments but does not include New Zealand superannuation.

20 page 16 Figure 3.12: Distribution by main source of income main source of income non-partnered individuals couples wages & salaries self-employment NZ superannuation other superannuation other income support investment income other regular or one-off income % Total income Figure 3.13: Proportion of non-partnered individuals and couples in each income band Income band Non-partnered individuals Couples $ % % Loss or zero , ,001-10, ,001-15, ,001-20, ,001-25, ,001-30, ,001-40, ,001-50, ,001-70, , , ,001 or more 2 14 The figure used for couples in figure 3.13 is the total combined income of the two partners. Three-quarters of non-partnered individuals reported their income as less than $30,000. In contrast, three-quarters of couples reported their total income as more than $30,000. A third of all couples had income greater than $70,000.

21 page 17

22 page 18 chapter four The distribution of net worth The 2001 Household Savings Survey (HSS) found that non-partnered individuals had an estimated net worth (assets less debts) of $ billion whilst the total for couples was an estimated $ billion. The distribution of this net worth within the two populations is discussed in this chapter. It considers the accumulation of net worth throughout life and the impact of age on net worth. Key concept Central to this report is the following key concept: Net worth of a couple is the combined net worth of both partners. Generally couples are treated as quite separate entities from nonpartnered individuals. But for some analysis the two groups are combined to form a population of economic units. Where analysis is done by individual characteristics (such as age, labour force status, ethnic group) the couple is assigned the characteristic of the member of the couple who was randomly selected to take part in the survey. For example, employed couples are those where the selected respondent was employed. Analysis Distribution characteristics Figures 4.1 and 4.2 show the distributions of net worth for non-partnered individuals and couples respectively. It is worth noting that in both cases the distributions are skewed to the right, and both are long-tailed, with the upper end of the data spread over a wide range of values. This illustrates the unequal distribution of net worth. Using figure 4.1 as an example, most non-partnered individuals have net worth under $50,001 whilst only a few have very high net worth.

23 page 19 Figure 4.1: Net worth distribution of non-partnered individuals number of people 450, , , , , , , ,000 50, and over net worth ($000s) Figure 4.2: Net worth distribution of couples number of couples 250, , , , , , ,000 75,000 50,000 25, ,050 1,050 1,125 1,125 1,200 1,200 1,275 1,275 1,350 1,350 1,425 1,425 1,500 1,500 1,575 1,575 1,650 1,650 1,725 1,725 1,800 1,800 1,875 1,875 1,950 1,950 and over net worth ($000s)

24 page 20 Likewise this skew is noticeable when considering the mean and median values. For all net worth estimates the mean value was higher than the corresponding median. For example: Figure 4.3: Mean and median net worth Economic unit Median net worth Mean net worth $ $ Non-partnered individuals 10,300 97,900 Couples 172, ,300 The median is used in this report as it is less affected by the extreme values than the mean. Life cycle accumulation Age distribution will impact on the accumulation and distribution of net worth. In the example above there appears to be a discrepancy between the net worth of couples and nonpartnered individuals. Even after sharing the net worth across the two partners, couples still have a median value over eight times that of their non-partnered counterparts. To a large extent this is due to the fact that net worth tends to accumulate over the life-cycle. For example 45 percent of year olds are in the lower quartile for net worth, whilst only two percent are in the upper quartile. So non-partnered individuals, which include a higher proportion of younger people (18-24 year olds make up 29 percent of all non-partnered individuals but only four percent of couples), also have a far lower median net worth. Figure 4.4 provides an overview of the distinct pattern of the distribution of median net worth by age groups for non-partnered individuals and couples. Figure 4.4: Median net worth by age median net worth $ non-partnered individuals couples 400, , , , , , ,000 50, total age group The importance of these life-cycle characteristics needs to be remembered when considering the overall distribution of net worth, or making comparisons between groups. Figures 4.5 and 4.6 show the proportions of each age group represented in each net worth quartile. Quartile 1 is the bottom 25 percent of the population when ordered by net worth. Conversely, quartile 4 is the 25 percent of the population with the highest net worth.

25 page 21 Figure 4.5: Proportion of age group in each net worth quartile for non-partnered individuals quartile 1 quartile 3 quartile 2 quartile 4 % age group Figure 4.6: Proportion of age group in each net worth quartile for couples quartile 1 quartile 3 quartile 2 quartile 4 % age group Analysing net worth distribution using deciles The population was divided into deciles to examine the distribution of net worth. That is to say, the population was ranked from lowest to highest net worth and then divided into 10 even-sized groups. Decile one is the 10 percent of the population with the lowest net worth whilst decile 10 is the 10 percent with the highest net worth. Figures 4.7, 4.8 and 4.9 show each decile plotted against their total net worth for three different groups: economic units; non-partnered individuals; and couples. In figure 4.7, the non-partnered individuals and couples have been combined to form one population of economic units.

26 page 22 Figure 4.7: Total net worth of economic units by decile total net worth $ billion decile Figure 4.8: Total net worth of non-partnered individuals by decile total net worth $ billion decile

27 page 23 Figure 4.9: Median net worth of couples by decile total net worth $ billion decile The uneven distribution of net worth is apparent in figures 4.7, 4.8 and 4.9. In considering the net worth of economic units (figure 4.7), the total value held by the lowest decile was -$3.303 billion, whilst it was $ billion for the highest decile. The largest absolute difference between deciles occurs between the ninth and tenth deciles, where the total net worth doubled. For non-partnered individuals the total net worth of the lowest decile was -$2.163 billion and the total net worth of the highest decile was $ billion (figure 4.8). The corresponding figures for couples were -$0.923 billion and $ billion respectively (figure 4.9). To draw conclusions about the implications of this data it would be necessary to consider the life-cycle effect. For example, figure 4.10 shows the equivalent net worth distribution for economic units in the year old age group. Because of the narrower age group, the life-cycle effects are cancelled to some extent. This allows a closer examination of the distribution of net worth within a particular group, in this case those approaching retirement. By this age the lowest decile has no total net worth rather than negative net worth, however, the distribution is still not equal. This is most noticeable when looking at decile 10 which is significantly larger than for other deciles.

28 page 24 Figure 4.10: Total net worth of year old economic units by decile total net worth $ billion The life cycle characteristics observed in this data may not be replicated over time. A longitudinal study or repeat cross-sectional surveys are needed to examine future life cycle characteristics. Net worth and age group To further examine the distribution of net worth by age, figure 4.11 shows the proportion of total net worth held by the top 20 percent of people in each age group. For example, of all economic units aged 35-44, the top 20 percent (in terms of net worth) had 67 percent of this group s total net worth. Figure 4.11: Net worth held by the top 20 percent of each age group Proportion Sum of net of net worth Total net worth of top held by top Economic unit Age group worth 20 percent 20 percent (billion) $ (billion) $ % Non-partnered individuals and over Couples and over Total economic units and over ,2,3,4 As mentioned in the following text, proportions greater than 100 percent occur due to the significant proportion of year olds with negative net worth

29 page 25 Note that there is an uneven distribution of net worth for every category, with the top 20 percent (in terms of net worth) accounting for around 60 percent of the net worth within most of the over-35 age groups. This suggests the spread of net worth is not only an attribute of differences in age. For the lower age groups the results are affected by the significant proportion of the population who have negative net worth. For example, 43 percent of non-partnered year old individuals had negative net worth. This high proportion of negative net worth meant that the top 20 percent of non-partnered individuals in this age group held 183 percent of this group s net worth. For this reason, it is useful to analyse those economic units with positive net worth, and those with negative net worth separately. Positive net worth of economic units Figure 4.12 orders only economic units with positive net worth into deciles. These represent 84 percent of all economic units. Here the ninth and tenth deciles hold nearly 70 percent of positive net worth, while by comparison the first and second deciles hold only 0.2 percent. Figure 4.12: Distribution of positive net worth and median net worth by decile for economic units Deciles Total net worth Median net worth % $ Decile Decile 2 0 4,800 Decile ,900 Decile ,800 Decile ,200 Decile ,800 Decile ,100 Decile ,900 Decile ,600 Decile ,400 Total ,500 Negative net worth of economic units Figure 4.13 provides a similar analysis for those economic units with negative net worth. In this case the sample is far smaller, so the analysis is by quintiles. Figure 4.13: Distribution of negative net worth by quintile for economic units Quintile Total negative net worth Median net worth % $ Quintile ,300 Quintile ,500 Quintile ,200 Quintile 4 5-3,000 Quintile Total 100-7,000

30 page 26 The lowest quintile of economic units with negative net worth holds 60 percent of the total negative net worth balance. This figure is the same for couples and non-partnered individuals when these groups are analysed separately. Negative net worth of non-partnered individuals and couples Separate analysis of the two populations that make up total economic units shows that an estimated 23 percent of non-partnered individuals and 8 percent of couples had negative net worth. Figures 4.14 and 4.15 show the distribution of negative net worth by age for non-partnered individuals and couples. The age group with the highest proportion of people with negative net worth is the age group for non-partnered individuals and couples. In comparison, only three percent of non-partnered individuals and one percent of couples aged 65 years and over, had negative net worth. The high proportions for those under 25 seem to reflect high levels of student loan debt plus relatively high levels of bank and credit card debt. Figure 4.14: Proportion of each age group with negative net worth, non-partnered individuals % age group Figure 4.15: Proportion of each age group with negative net worth, couples % age group

31 page 27 The highest proportion of people with negative net worth was non-partnered individuals and couples who were unemployed (by the labour-force status of the selected respondent). For non-partnered individuals, 46 percent of the unemployed and approximately 22 percent of those employed or not in the labour force had negative net worth. For couples, the pattern was similar with 36 percent of the unemployed and approximately seven percent of those employed or not in the labour force having negative net worth. High net worth Information collected showed that approximately 15 percent of all non-partnered individuals and 45 percent of all couples had net worth greater than $200,000. High net worth by age Figures 4.16 and 4.17 show the percentage of each age group (for non-partnered individuals and couples) with total net worth greater than $200,000. Figure 4.16: Proportion of each age group with net worth greater than $200,000, nonpartnered individuals % age group Figure 4.17: Proportion of each age group with net worth greater than $200,000, couples % age group The figures for year olds are not displayed in figures 4.16 and 4.17 as the estimates are not considered statistically reliable. The age group with the highest percentage of both non-partnered individuals and couples with net worth greater than $200,000 is the age group. The age group with the lowest proportion of non-partnered individuals and couples with net worth greater than $200,000 is the age group. Again this reflects the expected life cycle pattern of wealth accumulation.

32 page 28 High net worth and main source of income For non-partnered individuals and couples the median net worth was significantly higher for those who gave investment income as their main source of income. However only 2 percent of non-partnered individuals and couples received investment income as their main source of income over the year. For these non-partnered individuals, 69 percent had net worth greater than $200,000. For couples, the proportion was almost 100 percent. Those who gave other income support (this covers all Government benefits and allowances such as student allowances, family support and ACC earnings-related payments but does not include NZ Superannuation) as their main source of income were predictably the least represented, with 5 percent for non-partnered individuals and 16 percent for couples. A measure of inequality the gini coefficient One measure of inequality is the gini coefficient as calculated from a Lorenz curve 5.This is a measure often used to assess income inequality. The gini coefficient ranges between zero and one. The closer the number is to one the more unequal the distribution. The population of economic units in the HSS has a gini coefficient of indicating an unequal distribution. As a comparison, the gini coefficient for household disposable income in 1996 was This indicates there is more inequality in the distribution of net worth than in income. Going by this measure, the inequality in the distribution of net worth for non-partnered individuals was greater than for couples. The gini coefficients were and respectively. 5 Where the x-axis is the cumulative proportion of people and the y-axis the cumulative share of net worth. 6 Statistics New Zealand, New Zealand Now: Income (1998)

33 page 29

34 page 30 chapter five Ethnic groups The data from the 2001 Household Savings Survey (HSS) shows that accumulation and level of net worth is influenced by many factors. As well as a strong correlation with age, other socio-economic factors such as employment, education, ethnicity and marital status have an impact. These are investigated in this report. This chapter looks specifically at how different ethnic groups and cultural environments relate to the level of net worth. Concepts The HSS asked respondents (and partners where applicable) to list all ethnic groups they identified with. These were prioritised during data processing (see Appendix 1) so that each person was assigned one ethnic group. This makes analysis by ethnic group simpler. The prioritised ethnic groups are European/Pākehā, Māori, Pacific peoples, Asian and Other. The Other ethnic group is excluded from the analysis below as it covers a mixed range of responses that did not fit into the other categories. The analysis in this section is by economic units, that is the population of non-partnered individuals and the population of couples. Unless otherwise stated, couples were assigned the ethnic group and age of the partner randomly selected to take part in the survey. Analysis Age Non-partnered European/Pākehā individuals and couples had the highest net worth; followed by Asian, Māori and Pacific people (see figures 5.1 and 5.2). This is consistent with Māori and Pacific peoples being over-represented in the unemployed and low income groups which have lower net worth. Part of this difference is also an age group effect which is most obvious when looking at the population of non-partnered individuals. The non-partnered European/Pākehā individual population is significantly older than the other ethnic groups with approximately 22 percent of non-partnered individuals aged 65 or older. This compares with 6 percent of Māori, 7 percent of Pacific peoples and 4 percent of non-partnered individuals of Asian ethnicity. The higher proportion of middle-aged and older adults in the European/Pākehā ethnic group pushes up the median net worth which accumulates with age. Conversely, the Māori, Pacific peoples and Asian ethnic groups have large proportions of non-partnered individuals aged (61 percent of Māori, 62 percent of Pacific peoples and 68 percent of Asians). As negative or low net worth is prevalent in

35 page 31 the younger age groups, this age structure is one factor that contributes to the lower median net worth for these ethnic groups. Other contributing factors to differences in net worth by ethnic group are differences in education, labour-force status, income, and socio-economic status. Recent migration to New Zealand is also a factor for some of the groups. For example, when looking at the labour-force status of non-partnered individuals by ethnic group, HSS data showed that 2 percent of Europeans/Pākehā were unemployed while 8 percent of Māori and 13 percent of Pacific peoples were unemployed. Figure 5.1: Net worth by ethnic group (non-partnered individuals) Ethnic group Mean Median Mean/median $ $ ratio European/Pākehā 119,900 21,700 6 Māori 38, Pacific peoples 46,400 0 Asian 59,900 3, Total 97,900 10, Figure 5.2: Net worth by ethnic group (couples) Ethnic group Mean Median Mean/median $ $ ratio European/Pākehā 369, ,900 2 Māori 138,800 34,700 4 Pacific peoples 58,500 11,100 5 Asian 224, ,100 2 Total 322, ,900 2 Mean/median ratios Mean/median ratios are also shown in Figures 5.1 and 5.2. The mean, like the median, is a measure of distribution but is more influenced by extreme high and low values than the median. The mean/median ratio is one way of indicating how evenly net worth is distributed. The closer the ratio is to one, the more evenly spread the distribution. Note the ratio is not given for non-partnered Pacific peoples as it is not possible to divide by zero. Using this measure, figures 5.1 and 5.2 show net worth distribution is less equal within the Pacific peoples and Māori ethnic groups than the European/Pākehā and Asian ethnic groups. The large values for the mean/median ratios amongst non-partnered individuals reflect the effect of the Māori and Pacific populations having proportionally larger numbers of young adults, with low or negative net worth. The large differences seen amongst non-partnered individuals in all ethnic groups, except European/Pākehā, is not evident with couples. Instead, the mean/median ratios for net worth are similar to the European/Pākehā group. The distribution of net worth for European/Pākehā non-partnered individuals has two peaks, as shown in figure 5.3. This is called a bi-modal distribution. This reflects the differential net worth situations of younger non-partnered individuals compared to older non-partnered individuals, a group which is influenced by widowed, separated and divorced people with significant net worth.